Ask HN: Why it is seemingly hard to break duopolies in hardware? | Hacker News Reader

Ask HN: Why it is seemingly hard to break duopolies in hardware?

14 pointsspeeder8 comments

I noticed that for desktop Add In Board GPUs there are only nVidia, and AMD.

x86 CPU, Intel and AMD.

Ram chips, SK Hynix and Samsung (Micron and Elpida were tired in third/fourth place, when Elpida went bankrupt and got purchased by Micron, they temporarily overtook Hynix, but now their combined market share shrunk even further than before Elpida went bankrupt, cementing Hynix in second place).

Flat Panels I remember there were only a couple manufacturers, and just by Sony and someone else decided to stop making CRTs they killed the entire CRT market at the same time.

Internal Sound Cards, Creative and Asus Xonar.

Many other stuff is divided between Asus (ASMedia + spinoffs, like ASRock) and Formosa Plastics (owners of Via, WonderMedia, S3, Centaur...)

HDDs now are mostly WD and companies they purchased, or Seagate and the companies they purchased.

and the list keeps going on.

Many of these look ready to be disrupted by new competitors... but there are few attempts, for example PowerVR started as desktop GPU product, Intel tried to get into GPUs, Via still attempts sometimes to launch x86 CPUs, but for most part, all these markets are de-facto duopolies, and sometimes real duopolies, having literally no competitors (not even tiny ones).

Why is that?

CPU's, GPU's, RAM, Flat Panels, HDD's, etc require billions to even do something competitive, probably protected by many patents, are relatively stable industries with little market changing breakthroughs by new companies(one counter example is Intel with it's 3DXpoint memory, but it's Intel) partly because it takes tons of money to come up with them, have(i think) a lot of impact on their relevant supply chain.

On the other hand, if you're a researcher that came up with some innovation in those technologies, it makes more sense either to license or let your startup be bought by the big companies.

And on top of that we don't have a working anti-trust system, so companies just buy their way to duopolies.

Best guess - Network effects (of more than one kind but rooted in very low marginal costs to produce one more unit) tend to favor monopolies - but monopolies tend to keep prices very high until they have at least one competitor. Then they lower their prices. But the real reason may well be that the government lands much harder on monopolies with no competition - so it's actually more profitable in the long run to have one competitor with 20% of the market than to try to grab the whole market. High tech monopolies in particular have been know to boost or subtly subsidize these pet competitors to keep them around, and the AG's attack dogs away. I suspect we'll see more thorough regulation in the next few decades so this somewhat cheesy strategy might stop working.

The real duopoly is x86 vs ARM - Intel and AMD were a forced dual source, long ago. So why doesn't MIPS get more love? Does RISC-V have a chance? Intel's disinterest in low-powered chips opened up enough space for one competitor in a slightly different niche.

I think you described economies of scale, not network effects :)

Duopolies seem to be the natural outcome of any free markets (in the best case - in the worst case we get monopolies).

monopolies are illegal.

No they are not. Monopolies are legal.

What is illegal if they use (abuse) their market power to keep competitors out. For example AT&T had for the better part of a CENTURY a monopoly in the US. That didn't trigger the forced breakup. Instead:

> In the 1970s, the Federal Communications Commission suspected that the American Telephone & Telegraph Company was using monopoly profits from its Western Electric subsidiary to subsidize the costs of its network, which was contrary to U.S. antitrust law.

https://en.wikipedia.org/wiki/United_States_v._AT%26T_Co.

I wonder what PCs would be like if Intel bought Compaq in 1991 with people like Rod Canion and Jim Harris staying on. I think they were there when Compaq reverse engineered the IBM PC BIOS for example. Thinking about it, an alternative would be buying AST Research instead, which shut down in 1996 (just before ACPI was created) and was having trouble even beforehand. I assume that eventually AMD would do the same thing too.

This also reminds that Hynix had to be bailed out in 2001 and in 2003, US and EU imposed high tariffs as a a consequence. I wonder what would happen if Hynix actually failed.

Reply on news.ycombinator.com